Building a preventive maintenance program

A preventive maintenance (PM) program replaces surprise breakdowns with planned work. It is not a binder of manufacturer schedules, and it is not "service everything every quarter". It is a deliberate choice, asset by asset, about which failures are worth preventing, how, and how often — backed by records that prove the work was done and show whether it is working.

This guide walks you through building a PM program from the ground up: the asset inventory, criticality ranking, task design, scheduling, work orders, spare parts, backlog management, metrics and continuous improvement.

What a PM program is for

Preventive maintenance is scheduled work performed before failure to reduce the probability or consequence of failure. A good program:

• Reduces unplanned downtime on the assets that matter most.
• Makes maintenance workload predictable, so it can be planned and staffed.
• Extends useful life of equipment where wear is the main failure mode.
• Supports safety and compliance, with records of inspections and calibrations.
• Produces data for repair-or-replace decisions.

A bad program wastes effort on tasks that prevent nothing, disturbs equipment that was working fine, and generates paperwork nobody reads. The difference is in the design.

Step 1: Build the maintenance asset inventory

You cannot maintain what you have not listed. Start from your asset register and make sure every maintainable asset has:

• A unique ID and a tag, so technicians can open the record by scanning.
• Manufacturer, model and serial number.
• Location and responsible owner.
• Install or in-service date.
• Links to manuals, drawings and warranty documents.
• Its existing maintenance history, if any.

Group assets into asset types — for example, all identical compressors — so one PM plan can be applied to many assets.

Step 2: Rank assets by criticality

Not every asset deserves the same attention. A simple criticality score focuses effort where failure hurts most. Score each asset or asset type from 1 (low) to 3 (high) on:

• Safety and environmental impact of failure.
• Operational impact — does failure stop production, service or revenue?
• Cost of failure — repair cost, secondary damage, lost output.
• Redundancy — score 3 if there is no backup, 1 if a standby exists.

Illustrative scoring:

Assets scoring high get a full PM plan. Assets scoring low may be deliberately run to failure — a legitimate, documented choice when failure is cheap, safe and quick to fix.

Step 3: Understand how assets fail

PM only prevents failures that give warning or follow a predictable pattern. Ask, for each critical asset type:

• What fails? Belts, bearings, filters, seals, batteries, electronics.
• How does it fail? Gradual wear, contamination, fatigue, random electronic fault.
• Is there warning? Noise, vibration, heat, pressure drop, leaks.
• What does failure cost?

Wear-related failures (filters clogging, belts stretching, lubricant degrading) respond well to time- or usage-based tasks. Random failures (many electronic faults) are rarely prevented by servicing on a calendar — for those, inspections that detect early warning signs, or accepting failure with spares on hand, are often better strategies.

Step 4: Design the tasks

For each asset type, define the PM tasks. Start from the manufacturer's recommendations, then adjust for your operating conditions and failure history.

Each task should specify:

• What — "Replace intake filter" not "service compressor".
• Trigger — calendar (every 3 months), usage (every 500 operating hours), or both ("whichever comes first").
• Who — required skill or certification.
• Estimated time.
• Parts and consumables.
• Tools and safety requirements — lockout/tagout, permits, PPE.
• Acceptance criteria — measurable where possible, such as "belt deflection within manufacturer range".
• What to do if it fails — raise a corrective work order.

Types of PM task

• Inspection — look, listen, measure. Cheap and often the most valuable.
• Servicing — lubricate, clean, adjust, top up.
• Replacement — replace wear parts at intervals.
• Calibration — verify and adjust measuring instruments against a reference.
• Testing — function tests of safety devices and standby equipment.

Calendar or usage?

Usage-based triggers are more accurate for assets whose use varies — a generator that runs 50 hours one month and 400 the next. They require meter readings (hours, cycles, kilometres). Calendar triggers are simpler and suit assets used steadily, or whose degradation is time-driven (batteries, seals, fire equipment).

Step 5: Build the schedule

Turn tasks into a schedule:

1. Assign each task to assets and set the first due date.
2. Group tasks by location and trade, so one visit covers several tasks.
3. Level the workload — avoid all quarterly tasks falling in the same week.
4. Align with operations — schedule downtime-requiring tasks for planned shutdowns or low-demand periods.
5. Check capacity — is the total PM workload achievable with your team?

A worked capacity check

Illustrative example with assumed numbers.

A site has:

• 40 assets with a monthly 30-minute inspection: 40 × 0.5 h × 12 = 240 hours per year.
• 25 assets with a quarterly 2-hour service: 25 × 2 h × 4 = 200 hours per year.
• 10 assets with an annual 8-hour overhaul: 10 × 8 h × 1 = 80 hours per year.

Total planned PM: 240 + 200 + 80 = 520 hours per year, or about 10 hours per week (520 ÷ 52). If two technicians have about 60 productive hours per week between them, PM uses about a sixth of their capacity, leaving the rest for corrective work, projects and improvements. If the PM total exceeded available capacity, the program would need to be trimmed or staffed differently — a schedule that cannot be done is worse than a smaller one that is.

Step 6: Manage work orders

Every PM task, when due, becomes a work order. A work order should capture:

• Asset, task and due date.
• Assigned technician.
• Status — open, in progress, on hold (waiting for parts or access), complete.
• Actual time, parts used and cost.
• Findings and measurements.
• Follow-up corrective work raised.

Completed work orders are the maintenance history. They are what you will show an inspector, use to calculate reliability, and review when deciding whether to repair or replace.

Step 7: Plan spare parts

A PM task that cannot be done because the filter is out of stock is a failed PM task. For each PM plan:

• List the parts consumed per task.
• Calculate annual consumption (tasks per year × parts per task).
• Hold enough stock to cover lead time plus a margin.
• Identify critical spares — parts for critical assets with long lead times — and hold them even if rarely used.

Step 8: Control the backlog

Backlog is the work that is due or overdue but not done. Some backlog is healthy (it means work is planned ahead); too much means the program is not achievable.

• Review overdue PM weekly.
• Prioritize by criticality — never let critical-asset PM slip in favour of convenience tasks.
• Measure backlog in hours, not just count, and compare with capacity. For example, 120 hours of backlog with 60 productive hours per week is two weeks of work.
• Cancel or redesign tasks that are consistently skipped without consequence — they may not be needed.

Step 9: Measure the program

Useful metrics, all of which depend on good work-order records:

• PM compliance — share of PM work orders completed on time (for example, within a tolerance window around the due date).
• Planned maintenance percentage — planned hours ÷ total maintenance hours.
• Mean time between failures (MTBF) for critical assets — total operating time ÷ number of failures.
• Mean time to repair (MTTR) — total repair time ÷ number of repairs.
• Maintenance cost per asset and as a share of replacement value.
• Emergency work orders per month.

Example MTBF calculation: a pump runs 6,000 hours in a year and fails 3 times. MTBF = 6,000 ÷ 3 = 2,000 hours. If PM changes bring failures down to 1 the following year at the same run time, MTBF rises to 6,000 hours.

Calibration programs

Measuring instruments — gauges, meters, torque tools, scales, test equipment — need a specific kind of preventive task: calibration. An instrument that drifts out of tolerance can silently produce wrong measurements, and in regulated industries that can invalidate work done with it.

A calibration program needs:

• An instrument register — every instrument that affects quality or safety, with a unique ID and tag.
• A calibration interval for each, typically from the manufacturer or your quality system, adjusted by history. Instruments repeatedly found out of tolerance need shorter intervals.
• Traceable calibration — performed against reference standards, with a certificate recording the results.
• As-found and as-left results — whether the instrument was within tolerance before adjustment matters, because an out-of-tolerance as-found result means work done since the last calibration may need review.
• Status control — instruments that are overdue or failed should be clearly marked and withdrawn from use.
• Custody records — knowing who used an instrument, and when, makes it possible to trace affected work if it is later found out of tolerance. This is where check-in/check-out and calibration meet.

Documentation and compliance evidence

For safety-critical equipment, regulators, insurers and customers may ask you to prove maintenance was done. Good evidence is a by-product of good work orders, not a separate exercise:

• Every completed task has a record — date, technician, task performed, results, parts used.
• Records are attached to the asset, not filed by date in a folder, so the full history of any asset is retrievable in one place.
• Measurements are recorded as values, not just "OK", where the acceptance criteria are numeric.
• Certificates and inspection reports from external contractors are attached to the relevant work order.
• Overdue and missed tasks are visible, with reasons, rather than silently disappearing.
• Retention periods follow your regulatory requirements.

When an inspector asks for the service history of a specific pressure vessel or lifting device, you should be able to produce it in minutes by scanning the asset or searching its ID.

Step 10: Improve continuously

Review the program at least annually, and after every significant failure:

• Did a failure occur that PM should have prevented? Add or adjust a task.
• Are tasks finding nothing, repeatedly? Consider extending the interval.
• Are some assets consuming disproportionate maintenance cost? Consider replacement.
• Is condition data available? Consider moving from time-based to condition-based tasks for critical assets.

Common pitfalls

• Copying manufacturer schedules verbatim without adjusting to real conditions.
• Vague tasks — "check machine" is not a task.
• No link between PM findings and corrective work.
• PM treated as optional when the team is busy — which is exactly when failures happen.
• No records, so the program cannot be evaluated or evidenced.
• Too many tasks for the team, leading to permanent backlog and demoralization.

Where a CMMS fits

A CMMS (computerized maintenance management system) holds the asset list, PM schedules, work orders and history. For many organizations, maintenance is one part of a broader asset-management need that also includes tracking, custody, audits and depreciation — in which case a platform that combines them avoids maintaining two asset lists. The comparison in preventive vs corrective maintenance shows how to evaluate PM candidates financially.

PM in Asetavo

Asetavo's maintenance module keeps maintenance on the same asset records as the rest of your register. Work orders carry status, cost and parts, linked to the asset; the Maintenance tab shows each asset's full history; calibration is logged through work-order completion, with due tracking; and preventive schedules produce a due list of upcoming work. Automatic generation of work orders from PM schedules, and time-based maintenance-due alerts, are on the roadmap. Technicians open an asset by scanning its tag in the mobile app, and maintenance cost by asset appears in reports.

Maintenance is part of the Business plan — see pricing.